Sustainable forecasting of crude oil prices using explainable AutoML and ESG-driven market signals

Purpose Set within the context of sustainability-oriented financial markets, this study investigates the predictive power of financial, economic, and Environmental, Social, and Governance-related indicators, including investor sentiment indices, eco-friendly investment proxies, and cryptocurrencies, on West Texas Intermediate crude oil prices. Design/methodology/approach This study employed H2O AutoML to model the movements in the price of West Texas Intermediate (WTI) crude oil using Generalized Linear Models, Gradient Boosting Machines, XGBoost, Deep Learning, and Stacked Ensembles. The model incorporates macroeconomic indicators, oil supply-demand data, and financial market indices, with data spanning from July 1, 2020, to May 1, 2025. Model performance was evaluated using root mean square error and mean absolute error. Feature contributions were evaluated using SHapely Additive exPlanations, partial dependence plots, and individual conditional expectation curves. Findings Across models, gold emerges as the most significant predictor, followed by the Green Bond Index, the ESG Index, and the S&P 500, underscoring the collaborative impact of conventional and sustainability-linked factors. The model demonstrates strong predictive performance (RMSE ≈ 0.76), while indicators such as the Financial Stress Index and the S&P Global Clean Energy Index also exhibit notable explanatory power. In contrast, sentiment and cryptocurrency variables show a relatively limited impact. Research limitations/implications This research underscores the crucial role of green finance indicators in energy market forecasting, suggesting that market participants and regulators should consider integrating both economic and environmental factors, particularly in the context of climate risks and global sustainability targets. Practical implications This research provides investors, analysts, and policymakers with critical financial and ESG drivers of oil markets, enhancing decision-making through transparent and reliable explainable AutoML tools. Originality/value This study makes a significant contribution to the integration of automated machine learning techniques with explainable AI tools. It examines the role of traditional financial factors, ESG, and sentiment-related factors in crude oil price forecasting.